DISSERTATIONES SCIENTIAE MATERIALIS UNIVERSITATIS TARTUENSIS 2
DISSERTATIONES SCIENTIAE MATERIALIS UNIVERSITATIS TARTUENSIS 2
SERGEI VLASSOV Investigation of
nanoscale interactions and mechanical properties of nanostructures using quartz
tuning fork based real-time
measurements
The study was carried out in the Institute of Physics, University of Tartu The dissertation was admitted on April 29, 2011 in partial fulfillment of the re- quirements for the degree of Doctor of Philosophy (material science), and allowed for defense by the Council of the Institute of Physics, University of Tartu.
Supervisor: Dr. Rünno Lõhmus, Institute of Physics, University of Tartu Opponents: Ass. Prof. Ion Marius Sivebæk, Department of Mechanical
Engineering, Technical University of Denmark, Denmark Dr. Valdek Mikli, Tallinn University of Technology, Faculty of Chemical and Materials Technology, Centre for Materials Research
Defense: July 4, 2011, at the University of Tartu, Tartu; Estonia
This work has been supported by graduate school “Functional materials and processes” receiving funding from the European Social Fund under project 1.2.0401.09-0079 in Estonia.
ISSN 2228–0928
ISBN 978–9949–19–693–7 (trükis) ISBN 978–9949–19–694–4 (PDF)
Autoriõigus: Sergei Vlassov, 2011 Tartu Ülikooli Kirjastus
www.tyk.ee Tellimus nr. 351
CONTENTS
LIST OF ORIGINAL PUBLICATIONS ... 6
ABBREVIATIONS ... 8
PREFACE ... 9
1. QUARTZ TUNING FORK ... 11
2. ATOMIC FORCE MICROSCOPY – GENERAL DESCRIPTION AND TIP CONTAMINATION PROBLEM ... 21
3. NANOTRIBOLOGY AND MANIPULATION OF NANOSTRUCTURES ... 28
4. STRUCTURAL PROPERTIES OF NANOPARTICLES ... 32
5. AIMS OF THE STUDY ... 36
6. RESULTS AND DISCUSSION ... 37
6.1. Applications of QTF in mass, biological and chemical sensing (paper V and patent VII) ... 37
6.2. Real-time manipulation of gold nanoparticles inside scanning electron microscope with simultaneous force measurement (Paper I) ... 41
6.3. Real-time measurements of frictional and mechanical properties on ZnO nanowires inside SEM (Paper VI) ... 52
6.4. Crystal mismatched layers in pentagonal nanorods and nano- particles (Papers II and III) ... 58
6.5. Method of cleaning the tip of atomic force microscopy (Patent IV) . 60 SUMMARY AND CONLUSION ... 62
SUMMARY IN ESTONIAN ... 64
REFERENCES ... 66
ACKNOWLEDGEMENTS ... 71
PUBLICATIONS ... 73
LIST OF ORIGINAL PUBLICATIONS
I. Vlassov, S; Polyakov, B; Dorogin, L; Lõhmus, A; Romanov, A; Kink, I; Gnecco, E; Lõhmus, R (2011). Real-time manipulation of gold nano- particles inside a scanning electron microscope. Solid State Communi- cations, 151, 688–692.
II. Dorogin, L.; Vlassov, S.; Kolesnikova, A.; Kink, I.; Lõhmus, R.; Roma- nov, A. (2010). Crystal mismatched layers in pentagonal nanorods and nanoparticles. Physica Status Solidi B, 247 (2), 288–298.
III. Dorogin, L.; Vlassov, S.; Kolesnikova, A.; Kink, I.; Lõhmus, R.; Ro- manov, A. (2010). Pentagonal Nanorods and Nanoparticles with Mis- matched Shell Layers. Journal of Nanoscience and Nanotechnology, 10 (10), 6136–6143.
Patents:
IV. Invention: Method for cleaning the atomic force microscope tip and the sample; Owner: Estonian Nanotechnology Competence Centre, Univer- sity of Tartu; Authors: Sergei Vlassov, Ants Lõhmus, Rünno Lõhmus, Ilmar Kink, Jevgeni Šulga; Priority number: P200700031; Priority date:
12.06.2007 Submitted:
V. S. Vlassov, O. Scheler, M. Plaado, R. Lõhmus, A. Kurg, K. Saal and I.
Kink; Integrated carbon nanotube fiber–quartz tuning fork biosensor VI. B. Polyakov, L. Dorogin, S. Vlassov, I. Kink, A. Lohmus, R. Lohmus,
Real-time measurements of frictional and mechanical properties on ZnO nanowires inside SEM
VII. Invention: Method and device for measuring the chemical and biologi- cal analyte or viscosity and surface tension of the liquid; Owner:
Estonian Nanotechnology Competence Centre, University of Tartu;
Authors: Sergei Vlassov, Kristjan Saal, Rünno Lõhmus, Margo Plaado, Ants Lõhmus, Ilmar Kink; Priority number: P200900061; Priority date:
12.08.2009
Author’s contribution
Paper I: the author participated in development of experimental equipment, in experiments and data processing. Responsible for composing the manuscript.
Paper II and III: responsible for synthesis of nanoparticles. Participated in pre- paration of manuscript.
Paper V: the author is responsible for experimental equipment, all measure- ments and data processing.
Paper VI: author participated in development of experimental equipment, in experiments, and composition of manuscript.
Inventions IV and VII: the author is responsible for essence of invention, for claim formulation, for tenor, and for testing of the prototype.
ABBREVIATIONS
AFM atomic force microscopy BSA bovine serum albumin
CNT carbon nanotube
DSP digital signal processing DDS direct data synthesis
FET field-effect transistors fcc face centered cubic FIB focused ion beam
MEMS microelectromechanical systems NEMS nanoelectromechanical systems NP nanoparticle
NW nanowire PLL phase locked loop
SEM scanning electron microscope SET single electron transistor SPM scanning probe microscopy
TEM transmission electron microscope QTF quartz tuning fork
SNOM scanning near field optical microscopy SQUID superconducting quantum interference device
PREFACE
The thesis is related to several important research fields – nanotribology, nano- mechanics and biosensing – and can thus be considered multidisciplinary.
However, there are strong links between the topics and every part contributes to a common ultimate goal – investigation of the interactions, behavior and the mechanical properties of materials at nanoscale.
From the technical side, the “kingpin” connecting most parts of the thesis is the quartz tuning fork (QTF), considered here in terms of high-resolution sensing for various scientific applications. In the results section of the thesis, the use of QTF in combination with nanoporous carbon nanotube (CNT) fiber for the development of the novel method and appropriate device for sensing the mass of adsorbed media in liquid as small as nanograms is proposed. Its application in tracking adsorption rate of protein on nanotubes is successfully demonstrated. Another application of QTF presented in the thesis is the force sensor for real-time measurements of the frictional and mechanical properties of gold nanoparticles (NPs) and ZnO nanowires (NWs) manipulated on a flat surface inside the scanning electron microscope.
In addition to the experimental measurements, the nanoparticles are also treated theoretically. New mechanism of stress relaxation in pentagonal nano- rods and nanoparticles is proposed.
Moreover, a novel method of cleaning an atomic force microscopy tip – essential part of the sensor in the above-mentioned manipulation experiments – is elaborated.
For convenience, the connections between the research topics within the dissertation are presented schematically in Figure A.
The thesis is divided into 6 chapters. In the first chapter, the construction and operation principle of QTF are described. A brief introduction to the quartz crystal based sensing is given. The second chapter provides basic introduction to AFM and the tip contamination problem. The use of QTF in AFM is con- sidered. The third chapter is dedicated to the controlled manipulation of nanostructures and its importance in terms of nanotribology and nano- mechanics. In the fourth chapter, the main structural properties of gold nano- particles, which are treated both experimentally and theoretically in the results section, are presented. The fifth chapter defines the aims of the present study. In chapter six, the essential results are described.
The work was performed mainly at the Institute of Physics, University of Tartu. A significant contribution in development of biosensing method was added by the Institute of Molecular and Cell Biology, Estonian Biocentre.
Figure A. Connections between topics of the thesis
BACKGROUND 1. Quartz Tuning Fork
1.1. Structure and operation principle
1.1.1. Tuning Fork
The tuning fork is one of the best mechanical oscillators. It was invented in 1711 by the English trumpeter John Share. The important mode of the tuning fork is the one where the two prongs oscillate in a mirrored fashion [1]. This has the unique advantage that the center of mass stays at rest and all forces are compensated inside the material connecting the two prongs.
1.1.2. Quartz Tuning Fork – general information
Quartz tuning fork (QTF) (fig. 1a) is a fork-shaped quartz crystal with thin-film metal electrodes deposited on both sides of the QTF beams. Since quartz is a piezoelectric material, QTF can oscillate laterally under applied ac voltage.
Furthermore, the piezoelectric effect allows exciting and detecting the oscil- lation parameters (frequency, amplitude and phase) simultaneously.
The quality factor (Q-factor) is a fundamental quantity for characterizing the behavior of the resonator under the influence of external perturbing forces. It is defined as the ratio of the energy stored in the resonator to the energy loss during each oscillation period [1]. Due to the symmetrical shape and the fact that quartz is one of the materials with the lowest internal mechanical losses, Q- factor of QTF is very high – up to100 000 in vacuum and 10 000 in air. [2].
Assuming the properties of QTF to be isotropic and neglecting the influence of the electrodes, the resonance frequency of QTF can be estimated from its mechanical parameters from simple relation [3]:
(1.1.2.1) Here, I is the moment of inertia (I = ωt3/12 for rectangular cross-section), Eq is the Young’s modulus of quartz (7.87 x 1010 N m–2) and l, ω and t are the length, width and thickness of the cantilever, as shown in Fig.1b. meff is the effective mass for the oscillation beam, which is related to the real mass, m, through meff = 0.2429m [4]. The resonant frequency of the most commercially available QTFs is 32.768 kHz (215 Hz). More detailed theory is given in section 1.2.
a b
Figure 1. A photograph of the (a) encapsulated and (b) bare QTF; (c) a schematic diagram of the QTF beam.
QTF is mainly used as a frequency standard in various electronics and electro- mechanical systems. E.g. in a quartz watch, the fork is kept vibrating by an oscillator circuit which supplies voltage to the QTF electrodes. The quartz crystal itself is a capacitive component of that oscillator circuit. The alternating voltage of this circuit is detected, and then divided electronically to become a 1Hz signal, which is used to drive a stepper motor. Other applications are gyro- scopes, microbalances, and various sensors, discussed in more details below.
Due to the large industrial production QTFs are available at very low cost.
1.2. Theory of QTF
QTF is an electromechanical oscillator and its properties and behavior can be modeled using either electrical or mechanical approach. [5]. Main ideas are given below.
1.2.1. Electrical model
Piezoelectric oscillators can be modeled by an electronic equivalent circuit called the Butterworth-Van Dyke circuit (fig. 2) [6, 7]. The LRC resonator models the mechanical resonance: the inductance stands for the size of the kinetic energy storage, i.e., the effective mass, the capacitance reflects the potential energy storage, i.e., the spring constant and the resistor models the dissipative processes [8]. The parallel capacitance is given by the contacts and cables. The transfer function Y(ω)=I(ω)/U(ω), the so-called admittance is
(1.2.1.1) and is experimentally measurable. Due to the parallel capacitance C0 there are minimum in the admittance shortly after the maximum. On the resonance the current through the LRC branch flows in the phase with the voltage. The current through the parallel capacitance has a phase shift of 90 degree and causes a small phase shift of the total current. However, the admittance of the capa- citance C0 is small compared to the admittance of LRC branch and can be neglected or compensated electronically with a bridge circuit.
Figure 2. Butterworth-Van Dyke equivalent circuit for a piezoelectric resonator.
1.2.2. Electromechanical coupling
A fit of equation (1.2.1.1) to the experimental data works extremely well. Out of the electrical data the parameters L, R, C and C0 are obtained. These parameters are not sufficient to determine the mechanical oscillation amplitude. An addi- tional parameter is needed: the piezo-electromechanical coupling constant. It describes the charge separation Q on the electrodes on the piezomaterial per mechanical deflection x:[α]=C/m. With a simultaneous measurement of the electrical response and the mechanical amplitude with an optical interferometer this constant can be determined [8]. This constant is characteristic for one type of resonator and a modification, for example the attachment of an object or change of the environment will not alter this constant. The mechanical amplitude can be determined by the current I through the resonator:
(1.2.2.1)
To model the mechanical resonance, an energetically equivalent mechanical model consisting of one mass and one spring is applied (inset fig 3a). With the knowledge of the electromechanical coupling constant α, the mechanical
parameters can be determined from the electrical parameters by equating the potential energy Q2/2C = kx2/2 and the kinetic energy LI2/2= mv2/2:
, (1.2.2.2)
where k is a spring constant and γ is viscous friction.
From the electrical and mechanical correspondence the voltage can be iden- tified as the driving force: F = αU. Figure 4 shows the electric field in the crystal produced by the electrodes and how they are connected. This configu- ration detects and excites only movements of the prongs against each other. An interesting point to note is that there is a coupling of the two prongs via the piezoelectric effect. When one prong is deflected it produces a charge sepa- ration that in turn produces a voltage and thus deflects the other prong in the opposite direction. Assuming the QTF is not connected to any other electronics the charge is converted into a voltage over the capacitance C0 of the electrodes and the coupling constant is then α2/C = 57 N/m. However, if QTF is connected to cables, the capacitance is much larger and the coupling can be neglected. In the case of a fixed voltage (low impedance) the coupling is zero.
Figure 3. (a) Experimental measurement of the mechanical displacement of the front of a tuning fork prong at room temperature and a pressure of 10–6 mbar. The inset shows an energetically equivalent mechanical model (both prongs included). (b) The simulta- neous experimental measurement of the electrical response. The inset shows the para- meters for the electrical equivalent circuit.
Figure 4. Illustration of the electrical field lines in the cross-section of the QTF. The electric field along the horizontal direction causes a contraction or dilatation of the quartz in the direction perpendicular to the drawing plane. Only the movement of the two prongs in the mirrored fashion is electrically excitable or detectable.
1.2.3. Mechanical model
The mechanical model described in previous section energetically models the proper QTF mode and is appropriate for the determination of the oscillation amplitude. However, questions concerning asymmetries of QTF as they occur when preparing the QTF for the dynamic force detection (see chapter 2), cannot be answered with this model. A model that takes the two prongs into account has to be applied and is shown on the right in figure 5. This system has two modes, symmetric (in-phase) and antisymmetric (anti-phase), that are dege- nerate for vanishing coupling. The coupling splits the two frequencies and the two modes get mixed when the symmetry is broken. This model, however, cannot explain why the counter oscillating mode has a much higher quality factor than the synchronous mode, and thus is not the appropriate model for QTF. The model on the left in figure 5 has a third mass that models the move- ment of the base. In this model the counter oscillating mode still has a high quality factor because the center of mass stays at rest and all the forces are compensated inside the fork. The synchronous mode however, produces re- action forces in the support of the base and undergoes much stronger damping.
This model also explains the reduction of the quality factor when the symmetry is broken (e.g. by the attachment of additional mass to one of the prongs).
Examination of the model with the help of the Laplace transform and the influence of asymmetry is given in [5]. Numerical values obtained using this model are in a good agreement with experimental data. For example in figure 6 the calculated quality factor as a function of the additional mass is shown [5].
In conclusion it can be noted that the symmetry of the tuning fork is very important for a high quality factor. Any asymmetry lets the reaction forces act on the base of the tuning fork causing additional damping. Furthermore, other modes of the QTF do not generally interfere with the proper mode or have a much lower quality factor. This is in contrast to the model of two coupled oscillators, where the degeneracy is only slightly resolved by the coupling and the quality factors of both modes are approximately the same.
Figure 5: (right) Two coupled oscillators as a mechanical model for the QTF. (left) A model with a third mass to explain the influence of asymmetry on the quality factor.
Figure 6: The quality factor is reduced significantly when an additional mass is brought on to one of the prongs.
1.3. QTF as a sensor
Besides the main application as time standard, QTF is also a perfect device for making various sensors [2, 3, 9, 13]. Resonant frequency of the QTF is highly sensitive to some important parameters including added mass, density of surrounding media [9], as well as forces (load) acting on the crystal.
Currently, the most commonly used resonant sensors are plate-shaped quartz crystal microbalances (QCMs) with operation frequency ranging from one to several tens of MHz [10]. QTF is considered to be a cost-effective and simple alternative to QCM, as it has certain advantages. It has more stable resonant frequency. The electronics is simpler due to the considerably lower working
frequency. Moreover, lower frequency is preferable if measurements are made in liquid, since viscosity and other properties of the liquid can cause increased measurement uncertainties under high frequency excitation [11]. Unlike QCMs with thickness shear mode (TSM), QTF employs the flexural mode. Since the vibration amplitude of flexural modes is at least one level bigger than that of TSMs, the QTF is more sensitive to the external perturbations [9].
High sensitivity to mass loading is one of the most important properties of QTF in terms of sensing. The resonant frequency shift due to the mass loading on the QTF beams can be expressed by the simple equation [3]:
(1.3.1) where m and ∆m are the actual mass of QTF beam and the added mass, respectively. It should be noted that the formula is only valid in case of uniform rigid film covering both beams entirely.
The mass loading sensitivity of the tuning can be defined as:
(1.3.2) For the 32768 Hz QTF the sensitivity is in order of 10 ng/Hz [3], which is sufficient to sense monomolecular layers of material. This value is ~10 times smaller than for 10MHz QCM. However, for the QTF ~10 times higher counter accuracy can be employed due to lower working frequencies. In total, the sensitivity of QCM and QTF is of the same order of magnitude.
The ability to sense the small mass is used not only for direct mass measurements. The QTF can be coated or modified chemically to achieve the selective sensing. This idea is successfully applied in biosensing [2, 3]. For properly modified QTF frequency shift will take place only in case of specific bonding. It gives the ability to determine the presence of certain substances.
QTF can also be used to sense humidity if coated with appropriate water adsorbing film [12]. Additional mass originating from the adsorbed water will result in resonant frequency shift. The amount of adsorbed water depends on humidity.
Gas [9] and liquid [13] density sensors are another useful applications of QTF. On the basis of equations given by Zhang et al [3] the resonance frequency shift due to the density of the media can be written as:
(1.3.3) where n = 0.2429 is a constant [4], ρL and ηL are density and viscosity of the media respectively, ω0 is an angular resonance frequency.
In spite of all advantages, there is also a serious drawback that restricts the wide spread of QTF application in biosensing. Most of biological reactions take place in aqueous solutions. Moreover, in situ and real-time measurements are often demanded. Water is also most common solvent in chemistry. However, the arrangement of electrodes prevents the use of QTF in media with high dielectrical permeability like e.g. water. Thus, the choice of liquids for QTF measurements is limited to some organic solvents.
On more important application field for QTF is a force sensing. QTF is capable of sensing forces as small as e.g. interatomic interaction between the tip and the surface, and thus can be used as a sensor in Atomic Force Microscopy (AFM) [14]. More detailed description is given in next chapter where main principles of AFM are considered.
1.4 The Phase Locked Loop (PLL)
Since the QTF resonance is at a quite low frequency (33kHz), digital signal processing (DSP) can be applied. This is of great advantage because no analog devices can have a relative accuracy of 10–9 or better. As shown in figure 3 the phase between the excitation signal and the current through the fork as a function of the frequency is very steep at the resonance (about 180 degree/Hz).
This allows to detect any shifts in the resonance frequency very sensitively.
With a controller the excitation frequency is then automatically adjusted to maintain the phase at the value of the resonance. This is the idea of the PLL example of which is schematically shown in figure 7.
The deviation of the phase is detected with a digital two channel lock-in amplifier (SRS 830), which is synchronized by a digital signal from the fre- quency generator. The lock-in amplifier generates two orthogonal sinus signals as reference for the two channels. The phase of this reference signals with respect to the external synchronization signal can be shifted by an arbitrary value and is adjusted to have the x-reference signal in phase with the signal of QTF at the resonance. Ideally this phase shift would be zero (fig. 3), but the current-to-voltage converters and the long coax cables cause an additional phase shift. The output of the y-channel, which indicates any deviation from the resonance, is fed into controller to control the frequency to the resonance.
For very sensitive force detection parameters can be adjusted to achieve resolutions of the order of 1µHz. However, the stability of the reference frequency is specified to be 100µHz/oC and for the ultimate frequency shift detection an external reference frequency with a temperature controlled quartz oscillator or an atomic clock should be employed.
The oscillation amplitude of QTF is detected with the x-channel of the lock- in amplifier and the output signal is kept constant by a second feedback loop, which controls the amplitude of the excitation signal. This simplifies the interpretation of the different recorded signals, since the mechanical oscillation amplitude of QTF can be assumed to be constant. Second, the transients are
avoided that occur in response to a sudden change in the damping and could last up to seconds for high quality factors.
The PLL provides two signals that indicate the frequency shift and the excitation amplitude. Both signals can be used to control the probe sample distance by the z-feedback controller.
The power dissipated in the tuning fork is the product of the current and the voltage multiplied by the cosine of the phase angle between the two signals. The phase angle is zero on the resonance and is locked by the PLL. The current is kept constant by the amplitude controller and therefore the amplitude of the excitation signal is a direct measure for the power dissipation. Any additional damping caused by probe sample interactions can be detected very sensitively in this manner. Additional power dissipations of the order of 1fW can be detected [5]. This corresponds to an energy loss of 0.2eV per cycle of the tuning fork with typical oscillation energy of the order of 105eV (for 1nm amplitude).
Figure 7: An example of phase locked loop for measuring the frequency shift of QTF.
1.5. Fundamental limits for the force detection with QTF
Fundamental limits for the dynamic force detection with a QTF [15] can be considered applying the formalism introduced by Albrecht [16]. As a mechanical model the simple spring mass model shown on the right in figure 5 will be applied. While absolutely correct for the modeling of the proper QTF mode, the complications arising from having two prongs are avoided. For the parameters given in figure 8, the thermal white noise drive is 192fN/(Hz)1/2 at 300 K and 11fN/(Hz)1/2 at 1K. Multiplied with the transfer function for the mechanical model, the spectral thermal motion results and is shown in figure 8.
Assuming that the deflection detection can detect such small motions, the minimum detectable force gradient [15]
(1.5.1)
is 4.3mN/m at 300K for a detection band width of 100 Hz and an amplitude of 1 nm. The value gets smaller for low temperature (1K), smaller bandwidth (10 Hz) and larger amplitude (30nm): 3μN/m. However, experimentally this will be hard to realize, since the frequency shift that has to be detected is as low as 3μHz. This corresponds to a relative frequency shift of 10–10, which demands for a stability of the reference frequency that exceeds the values of standard equipment. To reach the thermodynamic limit with QTF, the deflection detection has to be able to detect the thermal noise off the resonance. For a current to voltage converter with a noise of 100fA/(Hz)1/2 the thermodynamic limit at room temperature could be reached with a detection band width of about 10 Hz. For the low temperature case, this is not possible with such a current to voltage converter. With a sensitive charge detector with a noise of 001e/(Hz)1/2, however, the thermal noise of the QTF at low temperature is dominant over a band width of about 50Hz. In conclusion it can be stated that to reach the thermodynamic limit for force detection, the detection bandwidth has to be narrowed or the quality factor has to be reduced to have the thermal noise of QTF dominant over the deflection detection noise. Experimentally one has to worry also about other sources of noise that could exceed the thermal noise of QTF. For example the noise of the excitation signal has to be smaller then 13nV/(Hz)1/2 which produces a force αU that corresponds to the thermal white noise drive of 11fN/(Hz)1/2 at low temperature.
Figure 8. The thermal motion of QTF at room temperature (300K) and at 1K. The motion is converted into a charge via the piezo-electro-mechanical coupling constant and into a current by multiplying the charge with the frequency.
It can be concluded that QTF is a powerful device for making various sensors having number of advantages over alternative devices including high q-factor, frequency stability, low cost, and elaborated models.
2. ATOMIC FORCE MICROSCOPY – GENERAL DESCRIPTION AND TIP CONTAMINATION PROBLEM
Atomic Force Microscopy (AFM) belongs to scanning probe microscopy (SPM) – the powerful morphological and structural analysis technique based on the sharp probe scanned over the sample surface, which has been employed in the analysis of a large range of materials with atomic resolution [17]. Appli- cation of AFM for manipulation of nanostructures and measurements of frictional and mechanical properties will be considered in the next chapter. The use of AFM tip in creation of QTF-based force sensor and appropriate AFM tip cleaning technique will be treated in results section. Here, a general description of AFM and QTF based AFM, tip contamination problem, and existing tip cleaning techniques are reviewed.
Since AFM is widely used in various research fields and its detailed descrip- tion can be found elsewhere [18], only the basic concepts are described.
2.1. Atomic Force Microscopy
The principle of AFM relies on the use of a sharp tip mounted on a cantilever which is brought into close proximity to the surface where intermolecular forces acting between the tip and the surface cause the cantilever to bend (Figure 9).
The tip is the scanned over the surface and images are obtained by recording the cantilever deflections during scanning detected with laser beam focused on the top of the cantilever.
Figure 9. AFM operation principle.
There are several operating mode in AFM technique. The general modes are contact mode, non-contact mode and dynamic contact mode (also called inter- mittent contact or tapping mode as patented by Bruker). In non-contact mode the attractive forces (generally the Van der Waals forces) are used to hold the tip above the surface. This prevents sample from harming. In contact mode the repulsive forces tend to dominate. The measurable forces in AFM technique are between 10–9-10–8 N. The figure 10 illustrates dependence of the force (F) between the tip and the sample on distance (R) between them.
Figure 10. dependence of the force (F) between the tip and the sample on distance (R) between them.
Tapping mode is the most commonly used of all AFM modes. Tapping mode imaging is implemented by oscillating the cantilever assembly at or near the cantilever's resonant frequency using a piezoelectric crystal. The piezo motion causes the cantilever to oscillate with a high amplitude (typically greater than 20nm) when the tip is not in contact with the surface. The oscillating tip is then moved toward the surface until it begins to lightly touch, or tap the surface.
During scanning, the vertically oscillating tip alternately contacts the surface and lifts off, generally at a frequency of 50,000 to 500,000 cycles per second.
As the oscillating cantilever begins to intermittently contact the surface, the cantilever oscillation is reduced due to energy loss caused by the tip contacting the surface. The reduction in oscillation amplitude or frequency is used to identify and measure surface features. Also, the phase shift between the input and output to the cantilever can be detected.
Tapping mode overcomes problems associated with friction, adhesion, electrostatic forces, and other difficulties that an plague conventional AFM scanning methods by alternately placing the tip in contact with the surface to provide high resolution and then lifting the tip off the surface to avoid dragging the tip across the surface. When the tip contacts the surface, the high frequency makes the surfaces stiff (viscoelastic), and the tip-sample adhesion forces is greatly reduced. Tapping mode inherently prevents the tip from sticking to the surface and causing damage during scanning. Unlike contact and non-contact
modes, when the tip contacts the surface, it has sufficient oscillation amplitude to overcome the tip-sample adhesion forces. Also, the surface material is not pulled sideways by shear forces since the applied force is always vertical. [19]
The use of modified probes enable much more specific information, such as frictional, magnetic, and thermal properties of the surface being investigated.
Different areas of a sample can cause different cantilever twisting depending on frictional forces acting between the tip and the surface. [20]
2.2. QTF in AFM
QTFs were introduced into SPM by Gunther, Fischer and Dransfeld [21] for use in scanning near field acoustic microscopy and later by Karrai and Grober [22]
and others [23], as a distance control for a scanning near field optical micro- scope (SNOM). In these microscopes the optical fiber tip is oscillating parallel to the surface resulting in shear force detection. Shear forces where then explicitly investigated using QTFs by Karrai and Tiemann [24]. QTFs with a magnetic tip were also used for magnetic force microscopy [25]. Rensen et al.
where able to resolve atomic steps with an atomic force microscope (AFM) [14]
cantilever and Si-tip attached to the QTF [26]. Giessibl et al. demonstrated atomic resolution on the Si (111)-(7x7) surface using QTF with one prong fixed (qPlus Sensor) [27]. Since only two electrical contacts are necessary for the operation, QTFs are simple to integrate in SPM even in a cryogenic environ- ment. The application of the QTFs for scanning probe microscopy at low temperatures was demonstrated by Rychen et al. [28]. Rozhok et al. [29] im- proved the construction of the sensor proposed in Rensen et al. [26] by gluing only the tip from AFM cantilever to one prong of QTF.
Main Advantage of QTF-based AFM over the conventional AFM are simple feedback electronics, compact design, and absence of optical detectors, making it well suited for use at cryogenic temperatures or inside electron microscopes for enhanced features like real-time observation of tip-substrate interaction etc.
Compared to micromachined AFM Si cantilevers the QTFs are very stiff.
The problems concerning the nonlinearity of the oscillator motion in the inter- action potential are reduced due to the high spring constant compared to the interaction forces. The stiffness avoids the snap in to contact and thus allows operating it with lower amplitudes then a cantilever. This simplifies the inter- pretation of the signals when the short-range interactions are investigated. The high stiffness is also of advantage for nanomanipulation applications as for example nano-lithography and manipulation. However, it is a disadvantage for the detection of very small forces, and is a danger for the tip to be crashed since the force is not limited by a soft spring.
QTFs are insensitive to high magnetic fields and operate well at low tempe- ratures. The fact that no light is needed for the deflection detection is important for the investigation of semiconductor heterostructures, which show the
persistent photo effect. Any light scattered on to the sample would alter its properties permanently.
Due to the high q-factor, dynamic force microscopy even in liquids is pos- sible. The perspective to use them as a carrier for sensorslike field-effect transistors (FETs), single-electron transistors (SETs), superconducting quantum interference device (SQUIDs) and hall sensors makes them attractive especially for the investigation of super- and semiconducting nanostructures in combi- nation with transport experiments.
The direct electromechanical coupling also allows to calculate the dissi- pation power, widely used in tribological measurements, easily and accurately without the troubles of calibration (U x I x cos(θ)). This is a very powerful advantage of piezo electric oscillators in dynamic force microscopy. [5]
2.3. AFM tips and tip contamination
The most common tip material in AFM techniques is silicon nitride (Si3N4).
This material is very hard which is necessary when the tip is dragged over the sample. It has outstanding wear resistance and good chemical resistance. Also, single crystal silicon can be used. It is somewhat less resistant to wear, but the tip can be made sharper, less than 10 nm as compared to 20–60 nm for silicon nitride. [30] Typical AFM cantilever with the tip is shown in figure 11.
Figure 11. Typical AFM cantilever with the tip (MikroMasch)
Ideally tip apex must be round shaped and terminate with the single atom.
However, scanning process or even exposure of AFM tip to ambient conditions can lead to morphological changes in physical profile of the tip and cause deviations from ideal shape and result in artifacts in the scanned image. There are a number of tip defects, which cause artifacts. The most common are:
a) Multiple peaks on the apex being atomic scale protrusions. Every peak during scanning contributes in the tip-sample interaction. In the simplest case of double peak apex the features on the sample surface look twin in the
scans. Absence of this defect is especially crucial when measuring single macromolecules.
b) Flattered apex. This results in lowering of resolution power. The sharpness of the apex tends to decrease during consecutive contact mode scans of the sample surface.
c) Non-spheric apex. Results in geometry distortion of sample features.
These changes are caused by two main mechanisms: material aggregation on the tip, and breakage, or wear, of some regions of the tip. The first mechanism, material aggregation, is the most frequent and can be regarded as an un- avoidable consequence of the scanning process in certain cases [31, 32]. The aggregated material, organic or inorganic, comes from the sample itself, or a contamination layer, always present on the surface of a sample exposed to ambient conditions. Even if great care is taken during the scanning process, loose particles may attach to the tip during its contact with the sample surface.
Consequently, tip effective dimensions are altered (enlarged), decreasing the image spatial resolution. Generally, material aggregation is the main cause of tip deterioration during investigations of soft samples (e.g. polymers and biological materials) and also brittle samples (e.g., oxides;) [31, 32]. Tips can be contaminated during the fabrication process or storage as well. Thus, even the new tip may require cleaning to increase the image resolution.
2.4. Common Tip Cleaning Methods
The general difficulty, concerned with cleaning of the SPM tip, is its small dimensions. The most of usual cleaning techniques will certainly destroy the tip.
In the present day several different physical and chemical methods are used to remove contaminations from the tip surface. The most frequently used are UV- ozone treatment [33-35], various wet processes, [36, 37], plasma etching [38], as well as all possible combination of existing methods [39].
2.4.1. UV-ozone treatment
UV-ozone treatment is the most common and is very efficient in removal of various hydrocarbons. As a pretreatment the tip placed in oxygen at atmospheric pressure and irradiated by UV-light of 185 and 254 nm. In these conditions, oxygen is excited to be ozone or radical so that some hydrocarbons are de- composed under UV irradiation by forming volatile molecules such as H2O and CO2. Then, the clean silicon surface is obtained by growing a thin oxidized layer on the tip surface by heating it in oxygen at low pressure with subsequent removal of this layer by annealing in ultra-high vacuum (UHV) condition [33]
or by reducing it in suitable acid solution [34]. Heating at high temperatures and for a long time to remove the oxide layer possibly makes the tip blunt, because formation of volatile molecule as SiO on a SiO2 surface needs a Si atoms supplied from the inside, causing corrosion of the Si tip.
2.4.2. Wet processes
There are lot of acid solutions were tested and some of them with certain limitations were successfully employed as a cleaning agents for SPM tip (so- called acid bathing). Among them are sulphuric acid, hydrochloric acid (HCl), hydrofluoric acid (HF), and sulfochromic acid solutions under a range of concentrations. HF solution is the most common. It is very efficient in removing large amounts of inorganic material, and, therefore, is considered an excellent cleaning option when only inorganic contamination is present. But it must be noted that HF solution is very aggressive to the tip and eventually damages it if long bathing times are employed. Besides cleaning effect HF treatment can make the tip sharper [33].
The cleaning efficiency of the HF solution for different types of inorganic contaminants may not be associated with a direct attack on the aggregated material. Rather it seems to be associated with the removal of the silicon oxide layer that always covers a silicon-made SPM tip, either new or used. It is well known in the semiconductor community that HF readily dissolves silicon oxide [40]. Removing an oxide layer, where the inorganic contaminants are attached, leads to cleaning of the SPM tip. This oxide removal may also explain the sharpening effect. However, such a general and, apparently, contaminant- independent mechanism is not very efficient when only organic material is aggregated to the tip. A possible explanation for this may arise from shielding effects of a compact organic layer covering the tip. While it is expected that hard inorganic materials cover the tip unevenly, forming holes where the cleaning solution can penetrate and dissolve the underneath silicon oxide layer, it is supposed that soft organic material may cover the tip evenly, producing a shield without holes, which precludes the penetration of the HF solution and, therefore, the organic contamination removal.
In contrast with HF solution, the cleaning efficiency of HCl solutions is not re- lated to silicon oxide dissolution, as it is insoluble in HCl [40]. Hence, it might rather be associated with a direct attack on the inorganic contaminating agent.
Despite the fact that cleaning with acids is fast and easy, other cleaning procedures are often preferred because of avoiding chemical damages to other parts of the cantilever including a piezoresistive film and a Si tip.
Less aggressive wet cleaning processes are based on the solvents like toluene [41], ultrapure acetone, tetrahydrofuran, ethyl alcohol, isopropyl alcohol, deio- nized water etc. Ultrasound can be noticeably helpful in the wet cleaning pro- cesses. With its help, even water (bi-distilled and de-ionized) can be used for tip cleaning [42]. Due to its strongly damaging action on many adhesive joints, water is a suitable liquid for cleaning by ultrasonic cavitation, dispensing with the use of any other cleansing agent, and it is recommended for the removal of particles in the micrometer size range, from solid surfaces. The fast periodic compression and decompression of high-surface tension liquid such as water produces a myriad of micro bubbles bursting within the liquid, especially at the existing solid-liquid surfaces. The resulting pressure gradients are sufficient to dislodge the particles, but they can also produce geometrical deformations at the surfaces [43]. The
cleaning time should be kept as short as possible, to avoid probe break-down. In some cases, the adherent particles could not be removed in a single step and cleaning procedure has to be repeated to achieve satisfactory results.
In the electronics industry to clean silicon and silicon nitride surfaces so- called piranha solution is used. It consists of sulphuric acid and hydrogen peroxide mixture. Dipping the tip in a piranha solution for 30 minutes can remove silicon oil contamination often introduced from the cantilever packing material [43].
2.4.3. Plasma Etching
Plasma etching, derived from the electronics industry, is known to remove organic contaminants from silicon and silicon nitride surfaces [43]. Usually hydrogen, oxygen or argon plasma is used. It reacts with carbon compounds or oxides on the tip surface.
2.4.4 Combined Methods
Sometimes it is reasonable to use different combinations of the existing cleaning methods. In this area, one of the most efficient combinations for organic removal involves UV/ozone exposure followed by ultrasonic solvent baths (15 min. in acetone and 30 sec. in isopropyl alcohol) [39]. The role of both solvents is only to help remove the fragments of organic material which were modified (chemical bonds broken), oxidized, and even vaporized by the long exposure to the ultraviolet radiation and ozone combination.
The combination of the sulfochromic solution with UV/ozone exposure and HCl acid solution bath can give satisfactory results in many cases [39].
Sulfochromic solutions constitute a well-known type of glassware cleaning agent. They are regarded as very efficient in removing organic and inorganic contaminants without damaging the glass surface. The most efficient sulfo- chromic solution had the following composition: 20% (v/v) of concentrated sulphuric acid, 17% (v/v) of potassium dichromate, and 63% (v/v) of demo- nized water. Furthermore, it was observed that the efficiency of the sulfo- chromic solution increases as the temperature increases and, thus, it should preferably be used at boiling temperatures.
In comparison with the specific cleaning processes, the combined methods are not as fast and not as simple. Therefore, it can be suggested that if there is only one type of tip contamination and its nature is known, it may be faster and simpler to employ the specific methods. On the other hand, if it is known that there are both organic and inorganic contaminants, or the nature of contami- nations is ignored, then it is more effective to employ the generic method.
The effectiveness of each method depends on the nature and amount of the tip contamination. There is also a serious risk of impairing the tip. Thus, the employment of a given cleaning procedure may or may not result in effective contaminant removal. Moreover, the existing cleaning methods require the tip to be taken out or at least moved from its position above the sample.
3. NANOTRIBOLOGY AND MANIPULATION OF NANOSTRUCTURES
3.1. Nanotribology
Recent endeavors to understand nanometer scale friction, adhesion, and wear, as well as the related possibilities to control them, have generated an inter- disciplinary scientific area – nanotribology, addressing pure and applied cutting- edge research topics with tremendous potential impact on technology and every- day life, including safety, economy, life quality, energy and material saving, towards a sustainable development.
The science underlying friction is a very long-standing problem. After centuries of scientific and technical development, friction and the related pheno- mena constitute a vast and interdisciplinary field. Understanding the complex processes occurring at the interface of two materials in relative sliding motion (the science of tribology) is central to pure and applied sciences, e.g. in studying plastic deformation and fracture development in a contact zone, as well as to many technological problems including lubrication, wear, fatigue etc. Espe- cially at the smaller microscopic scales, interfacial forces become dominant due to the increase in surface-to-volume ratio. In nanotechnology, for example, friction and adhesion are limiting factors that constrain performance and life- time of microdevices, such as magnetic storage systems, micro-/nano-electro- mechanical devices (MEMS/NEMS), and aerospace components. Durable low- friction surfaces, wear-resistant materials and coatings, as well as suitable liquid and solid lubricants are in demand for hi-tech applications.
Recent developments in experimental techniques, dominated by the atomic/
friction force microscopes (AFM/FFM), provided insight into the nature of interaction between materials in contact and relative motion at the micro and nanoscale. Based on these atomistic approaches many of the previous primary and historic questions about friction are being reconsidered and freshly answered, and more are emerging. Studies at molecular scales reveal frictional behaviors that are markedly different from those observed in macroscopic systems – empirical laws of friction no longer hold at the nanoscale. The nature of the elementary nanotribological mechanisms, which intimately relate friction, adhesion and wear, and even more the possibility to control them by external means is still in its infancy, and remains a formidable challenge. [44]
3.2. Nanomanipulation
Nanoscale manipulation experiments have two general purposes. On one side they enable investigation of the material frictional, mechanical and other properties at nanoscale. E.g. one of the most fundamental and still unsolved problems in nanotribology (the science of friction at nanoscale) is dependence of friction on real contact area. Its understanding is crucial for filling the gap between nanoscale and microscale friction. From the other side manipulation
experiments have practical aspect. Exact 2D positioning and assembly of nano- structures is essential for nanotechonological applications [45], like e.g. creation of nanoelectromechanical systems (NEMS), for applications in nanoelectronics, in digital information storage etc.
3.2.1. Manipulation of Nanoparticles
The most commonly used tool for the manipulation of nanostructures in general and nanoparticles in particular is the AFM, discussed in more details in chapter 2. Several different approaches have been applied in AFM manipulation strategies. In dynamic mode, particles can be moved during the scanning pro- cess when amplitude of the tip oscillations is increased above a certain threshold value. Estimation of frictional force is usually made on the basis of dissipated energy that is calculated from the phase shift as follows [46]:
(3.2.1.1) Increasing the scan rate above a certain value rather than increasing oscillation amplitude yields similar results [47]. Another approach consists in switching the feedback off during manipulation [48]. In this case, the tip pushes particles and oscillations are not essential for the manipulation process; cantilever deflection is recorded.
Particles can also be moved in contact mode. For example, Dietzel et al. [49]
introduced a so-called “tip-on-top” strategy. In this method, the tip is first positioned on top of the nanoparticle approximately at its center. The nano- particle then follows the tip motion. The measured torsional signal is directly proportional to the interfacial friction between the particle and the substrate.
AFM manipulations have certain limitations. First, there is no real-time visual feedback concerning the contact geometry or the particle position and behavior during manipulation (i.e., whether it is rolling or sliding). Only indirect conclusions can be drawn based on the shape of the force curves [50].
Additionally, many AFM experiments are made in ambient conditions, meaning that a considerable amount of water is present on all surfaces under in- vestigation, complicating the interpretation of forces.
Another problem is the “aging” of a sample exposed to ambient conditions, resulting in sticking of the particles to the substrate [51]. Sticking increases significantly with time. Given that AFM manipulation experiments are time consuming, adhesion can increase even within single experimental series.
To overcome these obstacles, manipulation experiments should be per- formed in a vacuum environment with real-time visual control.
3.2.2. Manipulation of Nanowires
Nanowires (NWs) – ultrafine wires having typical diameter in the range of 1- 100 nm and high aspect ratio [52]. Tribological studies of NWs are of high relevance from both scientific and technological point of view: NWs are now among most important objects in modern science and have number of promising applications in nanotechnology. Mechanical and electrical properties of NWs may be superior in comparison to corresponding bulk material [53]. NWs can be made from a wide range of materials, and can be metallic, semiconducting or insulating. Semiconductor nanowires made of silicon, gallium nitride and in- dium phosphide have demonstrated remarkable optical, electronic and magnetic characteristics (e.g., silica nanowires can guide light around very tight corners) [54]. NWs have potential applications in high-density data storage, either as magnetic read heads or as patterned storage media, and electronic and opto- electronic nanodevices, for metallic interconnects of quantum devices and NEMS [55]. Plenty of prototype devices based on NW were already demonstrated during last few decades including sensors, resonance-tunneling diodes, light emitting diodes, photodetectors, electromechanical devices, piezo- resistors and etc. [56-62].
Zink oxide (ZnO) is one of the materials commonly used in NWs pre- paration. ZnO NWs have some interesting properties, like e.g. electric-field mediated tunable photoluminescence with potential applications as novel sources of near-ultraviolet radiation [63]. ZnO NW can also be used to produce a p–n junction that serves as a diode [64]. ZnO NW gas sensors are also re- ported [65].
Considering that fabrication of NW-based devices requires precise control over positioning and subsequent behavior of the NWs, it is evident, that deeper understanding of NW-surface bilateral tribology mechanisms is crucial from applicative point of view.
Number of methods of investigation of mechanical properties of NWs or nanotubes (NT) have been developed. Ambient AFM can be used for vertical loading of a nanowire suspended over a hole or a trench to determine Young modulus and mechanical strength. This method was applied on Ge NWs and carbon NTs (CNT) [66, 67]. Elastic properties and mechanical strength of SiC NWs and CNT deposited on low friction substrate (MoS2) and pinned from one end by evaporated metal pads were measured using AFM lateral force regime [68].
Common method of NW’s Young modulus determination consists in finding the resonance frequency of a NW fixed from one end and placed inside scanning electron microscope (SEM) by sweeping the frequency of external excitation [69]. Another method is based on lateral bending of NW free end by pushing it with calibrated contact mode AFM cantilever, while NW second end is fixed on an edge of rigid substrate. Elastic deformation force is calculated from visual deformation of a NW and a calibrated AFM cantilever inside SEM.
Method was applied to investigate ZnO NWs [70]. Axial loading or stretching of NW glued between rigid substrate and calibrated AFM cantilever or between
two AFM cantilevers was applied for Si and B NWs, as well as CNTs Si, B NW and CNT [71-74]. Analogous axial tensile in situ tests performed on ZnO and Si NWs using MEMS-based nanoscale material testing stage inside transmission electron microscope (TEM) [75, 76]. Real-time force measurement during NW bending was performed by contact mode AFM inside SEM to measure Young modulus of vertically grown arrays of SnO2 NWs [77].
Only few works reported of measurements of kinetic friction of NW on flat substrate. Manoharan et al. examined the kinetic friction force during dragging of a ZnO NW parallel to its axis at different loading forces measured by MEMS force sensor at ambient conditions [78]. Conache et al. reported distributed static and kinetic friction of InAs NWs on Si3N4 coated Si wafer based on measuring curvature of ultimate NW bending radius after AFM manipulation at ambient conditions, where friction was calculated using Young modulus of a bulk material for calculations [79]. In other words, existing experimental methods and theoretical models for measuring and description of tribological properties of NWs contain significant uncertainties and do not include all important parameters. Thus, it can be concluded, that NW research is in preparatory phase and its potential is not realized.
4. STRUCTURAL PROPERTIES OF NANOPARTICLES
Nanoparticles (NPs) are often considered as a fifth state of matter. They have some properties, unlike either the molecules or the bulk and in that sense represent a transition between those states. Many effects can appear in nano- particles, which are not observed in the bulk crystals. One of the most common NP material used in nanoscience and nanotechnology in general and in nano- manipulation experiments in particular, is gold (Au). Au NPs were thorough investigated both theoretically and experimentally in present thesis. This section provides necessary background of some particular properties of NPs in general and Au NPs in particular, like shape and surface energy, which are essential in the context of nanomechanics, nanotribology and nanomanipulation. Gold belongs to elements with face centered cubic lattice (fcc) and its structural properties can be considered in terms of fcc materials in general.
4.1. Structural Properties
Important property of NPs is that their shape and crystal structure differs from that of the bulk material. The main reason is the nature of the forces such as surface tension, acting on NP [80]. The most frequently observed shapes at the nanometric scale are shown in Fig. 12. They correspond to the cuboctahedron, the icosahedron, the regular Bagley decahedron, the star particle, the Marks truncated decahedron, and the round pentagonal configuration. The latter shape has several variants such as the truncated octahedron or the tetrakaidecahedron.
The regular decahedron, the star, the round pentagonal, and the Marks decahe- dron are variants of the decahedron shape and correspond to one of the most important shapes because they are very stable and frequently observed.
Particles of fcc materials have many variants of the basic shape and correspond to different truncations of the cuboctahedron. The most commonly observed are the truncated octahedron and the tetrakaidecahedron. For the case of gold and silver, all the forms of the truncated decahedron and the icosahedron become the most favorable shape. Another variant of fcc particles is the tetrahedral particle, which in a truncated version becomes flat platelets.
Figure 12. Main types of particles which are observed at the nanometric scale corres- ponding to (a) fcc cuboctahedron, (b) icosahedron, (c) regular decahedron, (d) star decahedron, (e) Marks decahedron, and (f) round decahedron.
Faceting and truncation are the most favorable mechanisms chosen by nature for minimizing the total energy of the particle. This is valid even for the fcc shapes in which a pure cubic particle has never been observed. The formation of extra facets induces the reduction in the contributions to the energy coming from the surface area and from the radius of curvature of the particles. As discussed by Cleveland et al [81] and Patil et al. [82] using macroscopic concepts as a guide, a particle at 0 K should have a total energy (Et) given by
Et(N) = EBN + Eσ + Eγ , (4.2.1) where N is the number of atoms, EB is the bulk energy per atom, Eσ is the strain energy per atom, Eγ is the average surface energy per unit area, and S is the surface area of the cluster. Faceting introduces a minimization of the second and third terms of Eq 4.2.1. Therefore, particles will tend to shapes containing extra facets and to an overall more rounded shape. However, from the point of view of atomistic simulation, the situation is more complex. As the size increases, the internal stress becomes very important, causes some particles such as the icosahedron to increase its energy very rapidly, and becomes less stable despite having the most rounded shape.
As the particles grow larger, they start to produce more internal stresses.
This can be considered as a slow transition to the bulk state. Therefore, a stress release mechanism should dominate at a given size. This produces more
complicated structures with various defects. This problem, for the case of decahedral particles, has been discussed theoretically in an extensive way by Gryaznov et al. [83, 84] Those authors suggested several mechanisms for stress release. From their calculations using classical theory of elasticity they conclude that three most energetically favorable mechanisms for stress release are:
dislocation formation, formation of a system of thin twin parallel layers in one of the sections of the decahedron, splitting of the pentagonal axis in two or more partial disclinations, and displacement of the pentagonal axis to the periphery of the particle. In general, several release mechanisms are acting at the same time and nanoparticles of a size >10 nm have a complex structure.
It is also important to know the value of the total energy versus the size and structure of the nanoparticles. Some simulations were performed using Lennard- Jones interactions [85] or more sophisticated potentials such as the Finnis- Sinclair and combined Lennard-Jones potentials and three body potentials. [86]
The most comprehensive are the ones of Landman et al. [81, 87] Contrary to earlier calculations that consider the icosahedron the most stable structure, Landman’s group concluded that one of the most stable structures in some sizes corresponds to the Marks decahedron. However, these calculations consider a relatively small number of atoms.
Yacaman et al [88] have calculated the energy of different structures as a function of the size for wide range of sizes and found that the most stable structures correspond to the truncated decahedral structures: the star, the rounded pentagonal, and the Marks decahedron. Although, for a small number of atoms, the icosahedrons and the regular decahedron are also more stable than the fcc structures. However, when the number of atoms increases, the truncated decahedral structures remain stable over the fcc whereas the icosahedron and the regular decahedron become less favorable. It should be noted that energies are so close that in a given sample it is expected that a statistical distribution of shapes will be observed specially for the case of smaller sizes. In addition, rapid growth conditions in which the particles tend to be out of equilibrium will lead to a more diverse distribution of particles. This is the case for vapor deposition growth. For colloidal methods, which produces slower growth and allows the particles to reach an equilibrium configuration the truncated decahedral shapes should be the predominant structure. This latter fact is confirmed by experi- mental observations. This is also true for postdeposition annealed particles. [88]
An important case to consider is when the particles are passivated with an organic molecule. In that case, an extra term in the energy is introduced by the interaction between the atoms of the NP and the atoms of the organic molecule.
This results in an even higher tendency to produce faceting shapes such as all the forms of the truncated decahedron or the most truncated fcc shapes as shown experimentally by Gutie´rrez-Wing et al. [89]
The situation with the particle shape is complicated even more by the fact that at nanoscale the shape is not necessarily constant. This is because the energy of a nanoparticle shows many local minima configurations, cor- responding to different structures. A small excitation (e.g. by the beam of an
electron microscope) may be sufficient to induce shape transitions on the particle like was shown e.g. by Ijima and Ichihasi [90] for a gold particle of
~2 nm size fluctuating between the cubo-octahedral, icosahedral, and single twined structure. In order to explain the structural fluctuations two main models have been proposed. The first one involves the complete melting of the particle followed by a recrystallization to a new structure [91]. The energy for the melting is provided by the inelastic scattering of the incoming electrons on the particle. The second model [92] assumes that different particle configurations have similar energies and the low energy barrier between different configu- rations allows transition without melting. This phenomenon has been termed quasimelting by Marks and co-workers [92] and refers to the fluid-like behavior observed in the NPs. In some cases, the particle has a transition with a memory of the original crystalline orientation [93]. The latter case involves the appearance, movement, or disappearance of twin boundaries.
Variations in shape and the presence of defects have strong influence on mechanical and tribological properties. In particular, the issue of understanding how friction operates at the nanoscale level is one of the most important issues in nanotechnology. This subject is still an open question and the phenomenon is very complex and requires extensive additional research.
5. AIMS OF THE STUDY
The ultimate goal of the study is contribution to detailed understanding of the interaction mechanisms and properties of materials at nanoscale, and finding critical parameters controlling them. To achieve this goal a number of objectives were assigned, which include both experimental and theoretical aspects. All objectives are interconnected and reinforce each other. The main objectives are listed below.
Elaboration of the nanomanipulation technique inside a SEM chamber, for real-time measurements of the frictional and mechanical properties of nanostructures (nanoparticles and nanowires).
Finding relations between morphology and frictional properties for gold nanoparticles manipulated on oxidized silica substrate inside SEM.
Development of the theoretical model of stress relaxation in gold nanoparticles and nanorods based on the formation of shell layer with crystal layer mismatch.
6. RESULTS AND DISCUSSION
6.1 Applications of QTF in mass, biological and chemical sensing (paper V and patent VII)
In this section the mass sensitivity of QTF is used to elaborate the chemical and biological sensing method for measurements in water, which is impossible with the existing QTF sensors for the reasons described in chapter 1. In all experi- ments described in chapter 6.1, QTF was driven electrically by AGILENT 33120A Function/Arbitrary Waveform Generator. The frequency response was tracked with a METRIX 3354 oscilloscope. METRIX software was used to record and analyze the data on a PC.
6.1.1. Controlled silanization in vapor
Before development of the QTF sensor for measurements in water, we per- formed simple experiments with unmodified bare QTF in gaseous atmosphere to investigate its sensing capabilities and test our equipment. Silanization of QTF in gaseous phase was tracked.
QTF, connected to electronic circuit, was placed in hermetically closed vessel and the resonance frequency of QTF was recorded. Then the drop of tetramethoxysilane was placed at the bottom of the vessel and the frequency response of QTF was continuously recorded. During 100 minutes the total resonance frequency shift was 27 Hz (fig. 13). According to eq. 1.1.2.1 it corresponds to 270 ng of added mass. To ensure that silanization and not just condensation of silane took place, the QTF was rinsed in acetone and frequency response was measured again. No restoration of resonance frequency was observed, indicating that silanization was complete.
Figure 13. Frequency spectra of vapor phase silanization of QTF. Frequency response corresponds to: 1 – clean QTF in air, 2 – QTF after 100 minutes in silane vapor atmo- sphere, 3 – QTF after rinsing in acetone and drying. Resonant frequency shift between 1 and 3 is approximately 27 Hz.
6.1.2. Insulation of QTF electrodes
We have tried to insulate the QTF electrodes to perform the measurements directly in buffer solutions, as was suggested by Su et al. [2]. We have tested different coating methods, including silanization, TiO2 atomic layer deposition, sol-gel and polymer coatings. However, due to the arrangement of the QTF’s electrodes, it acts as a capacitor, meaning that if the surrounding medium has a high dielectric permeability, there will be unavoidable capacitive losses, regard- less of the presence of the coating.
6.1.3. Integrated carbon nanotube fiber–quartz tuning fork biosensor
In this section, a novel label-free biosensor for in-situ measurements in aqueous solutions is described. The sensor is comprised of a carbon nanotube (CNT) fiber attached to one prong of a QTF. The CNT-fiber was chosen because of its porous structure, low density and high stiffness [94]. Moreover, CNTs can be easily functionalized [95], which is necessary for designing specific bio- recognition assays. The performance of the sensor was demonstrated experi- mentally by monitoring the adsorption rate of bovine serum albumin (BSA) to the CNT-fiber at two different pH values. According to Valenti et al., BSA is adsorbed directly onto CNTs, with the adsorption rate depending on pH [96].
For the BSA adsorption experiments, pH values corresponding to maximal (pH 4.8) and minimal (pH 7) adsorption rates were selected.
BSA was purchased from PAA Laboratories GmbH. BSA solutions (0.1 mg mL-1, pH 4.8 and pH 7) were prepared in phosphate buffers. Multiwall carbon nanotubes (O.D.×I.D.×L = 20–40 nm × 5–10 nm × 0.5–50 μm) were purchased from Sigma-Aldrich. Fibers were prepared by dielectrophoresis [97]. Briefly, the tip of a sharp tungsten wire was immersed into a droplet of a CNT suspen- sion in water and a fiber of desired length was drawn under an AC voltage. Our set-up enables preparation of fibers up to 100 mm in length. By changing the concentration of the CNT suspension, the drawing speed, and the voltage, it is possible to vary the diameter of the fibers from submicron to 400 μm. A typical CNT-fiber used in our BSA adsorption experiments is shown in Fig. 14.
To assure the comparability of individual sensors, a precursor fiber several centimeters long was drawn and then cut into equal length pieces to ensure uniformity of diameter and other parameters within the set of different sensors.
Biosensors were made by gluing the CNT-fiber to one prong of the QTF (Fres=32768 Hz, Clock quartzes TC-38) using an epoxy glue (Eccobond 286, Emerson & Cuming).
